Differential synthetic aperture radar interferometry (D-InSAR) is characterized mainly by high spatial resolution and high accuracy over a wide coverage range. Because of its unique advantages, the technology is widely used for monitoring ground surface deformations. However, in coal mining areas, the ground surface can suffer large-scale collapses in short periods of time, leading to inaccuracies in D-InSAR results and limiting its use for monitoring mining subsidence. We propose a data-processing method that overcomes these disadvantages by combining D-InSAR with the probability integral method used for predicting mining subsidence. Five RadarSat-2 images over Fengfeng coal mine, China, were used to demonstrate the proposed method and assess its effectiveness. Using this method, surface deformation could be monitored over an area of thousands of square kilometers, and more than 50 regions affected by subsidence were identified. For Jiulong mine, nonlinear subsidence cumulative results were obtained for a time period from January 2011 to April 2011, and the maximum subsidence value reached up to 299 mm. Finally, the efficiency and applicability of the proposed method were verified by comparing with data from leveling surveying.